Wikidata:Property proposal/probability density function
Jump to navigation
Jump to search
probability density function[edit]
Not done
| Description | Mathematical density function for a probability distribution |
|---|---|
| Represents | probability density function (Q207522) |
| Data type | Mathematical expression |
| Template parameter | "pdf" in en:template:Probability distribution |
| Domain | probability distribution (Q200726) |
| Allowed values | LaTeX |
| Example | normal distribution (Q133871) → "\frac{1}{\sigma\sqrt{2\pi}}\, e^{-\frac{(x - \mu)^2}{2 \sigma^2}}" |
| Robot and gadget jobs | This could potentially be filled from the templated in the English Wikipedia. |
- Motivation
To celebrate the arrival of the new data type "Mathematical expression" Finn Årup Nielsen (fnielsen) (talk) 13:30, 10 February 2016 (UTC)
- Discussion
Comment We should convert TeX string (P1993) and keep this property for all mathematical formula. Snipre (talk) 15:43, 10 February 2016 (UTC)
- As noted for another suggestion, the "Mathematical expression" will render the LaTeX. — Finn Årup Nielsen (fnielsen) (talk) 10:28, 11 February 2016 (UTC)
- Comment Note there is not "one" formula for a probability distribution (Q200726). There are mean, variance, PDF, CDF. Another approach could be to use a generic 'formula' property and then use qualifiers.
Support also "TeX string" probably should be rendered as a string given its name - and for example should not be converted to MathML as is apparently planned for some renderings of "Mathematical expression". Distinct properties for mathematical expressions on an item are definitely appropriate as this proposal (and others) makes clear. ArthurPSmith (talk) 13:29, 11 February 2016 (UTC)- Support that would be cool. --18:22, 11 February 2016 (UTC)~---
Question @Fnielsen, Snipre, ArthurPSmith: Is this property still needed? I see that normal distribution (Q133871) uses defining formula (P2534).--Micru (talk) 14:33, 4 March 2016 (UTC)
- Good question. :-) The formula shown is for the PDF, while every distribution has multiple formula, CDF, mean, variance, other moments, characteristic function, so what should we do about that? Have a whole series of properties? Or use qualifiers? If the equation is to be used in an infobox I suppose the qualifier approach is not well suited? — Finn Årup Nielsen (fnielsen) (talk) 21:50, 4 March 2016 (UTC)
- (To avoid confusion regarding "The formula shown is for the PDF": I have now added a second notation for a normal distribution (Q133871)) — Finn Årup Nielsen (fnielsen) (talk) 21:56, 4 March 2016 (UTC)
- @Fnielsen, Micru: I'm not sure this is really needed now. What might be needed are formulas for other properties of a distribution - for example if we wanted to create a "mean of probability distribution" property, or "variance", etc. But otherwise I think "defining formula" would be exactly what you would expect to see as a description of the distribution itself for, for example, logistic distribution (Q589603), binomial distribution (Q185547), Poisson distribution (Q205692) etc. ArthurPSmith (talk) 14:50, 11 March 2016 (UTC)
- @Fnielsen: I would have expected this ⟨ normal distribution (Q133871)If we use defining formula (P2534) that way, this proposal (and the next one) is unnecessary? However, this seems to stretch the original intention of defining formula (P2534), where "a theorem or a law" was to be represented. But of course, in broader sense any formula could be seen as representing a theorem or a law. Lymantria (talk) 10:14, 25 April 2016 (UTC)
⟩ defining formula (P2534) ⟨ \frac{1}{\sigma\sqrt{2\pi}}\, e^{-\frac{(x - \mu)^2}{2 \sigma^2}} ⟩
of (P642) ⟨ probability density function (Q207522) ⟩ - As no reply has reached this in a month:
Oppose - unneeded. Lymantria (talk) 15:01, 30 May 2016 (UTC)
- @Fnielsen: I would have expected this
- @Fnielsen, Micru: I'm not sure this is really needed now. What might be needed are formulas for other properties of a distribution - for example if we wanted to create a "mean of probability distribution" property, or "variance", etc. But otherwise I think "defining formula" would be exactly what you would expect to see as a description of the distribution itself for, for example, logistic distribution (Q589603), binomial distribution (Q185547), Poisson distribution (Q205692) etc. ArthurPSmith (talk) 14:50, 11 March 2016 (UTC)
- (To avoid confusion regarding "The formula shown is for the PDF": I have now added a second notation for a normal distribution (Q133871)) — Finn Årup Nielsen (fnielsen) (talk) 21:56, 4 March 2016 (UTC)
- Good question. :-) The formula shown is for the PDF, while every distribution has multiple formula, CDF, mean, variance, other moments, characteristic function, so what should we do about that? Have a whole series of properties? Or use qualifiers? If the equation is to be used in an infobox I suppose the qualifier approach is not well suited? — Finn Årup Nielsen (fnielsen) (talk) 21:50, 4 March 2016 (UTC)
- Oppose per Lymantria. --Izno (talk) 12:11, 31 May 2016 (UTC)
- Can be deleted for me. Snipre (talk) 13:54, 31 May 2016 (UTC)
- as I noted above, I don't think this is needed. Feel free to close as not done. ArthurPSmith (talk) 20:36, 9 June 2016 (UTC)
@Fnielsen: 
Not done, per discussion. defining formula (P2534) seems sufficient. --Srittau (talk) 23:41, 11 June 2016 (UTC)